Crypto Newbie / Simulators / Blob Fees (EIP-4844)
EIP-4844 Blob Fee Market Simulator
On March 13, 2024, the Dencun upgrade went live on Ethereum mainnet. Within 24 hours, L2 fees on Arbitrum, Optimism, and Base dropped 5-100×. The reason: EIP-4844 introduced BLOBS — a separate, much cheaper data type for L2 rollups to publish their batch data. This simulator runs blob fee market dynamics over 15-50 blocks under different demand patterns, then shows the rollup batch economics that gets you to $0.01-0.05 per L2 transaction.
Blob fee simulation
Result over simulated blocks
Final blob base fee
0.1229 gwei
Final cost per blob
$0.0564
Block-by-block
| Block | Blob base fee (gwei) | Blobs used | vs target (3) | Cost per blob (USD) |
|---|---|---|---|---|
| #1 | 1.0000 | 2 / 6 | 67% | $0.4588 |
| #2 | 0.7412 | 2 / 6 | 67% | $0.3400 |
| #3 | 0.5494 | 3 / 6 | 100% | $0.2520 |
| #4 | 0.5494 | 2 / 6 | 67% | $0.2520 |
| #5 | 0.4072 | 2 / 6 | 67% | $0.1868 |
| #6 | 0.3018 | 3 / 6 | 100% | $0.1384 |
| #7 | 0.3018 | 2 / 6 | 67% | $0.1384 |
| #8 | 0.2237 | 3 / 6 | 100% | $0.1026 |
| #9 | 0.2237 | 2 / 6 | 67% | $0.1026 |
| #10 | 0.1658 | 3 / 6 | 100% | $0.0761 |
| #11 | 0.1658 | 2 / 6 | 67% | $0.0761 |
| #12 | 0.1229 | 3 / 6 | 100% | $0.0564 |
| #13 | 0.1229 | 3 / 6 | 100% | $0.0564 |
| #14 | 0.1229 | 3 / 6 | 100% | $0.0564 |
| #15 | 0.1229 | 3 / 6 | 100% | $0.0564 |
Rollup batch economics (one blob)
L2 users packed per blob
1,638
Cost per blob
$0.0564
Blob cost per L2 user
$0.000034
Pre-4844 L1 calldata cost per user
$0.100800
Savings vs L1 calldata
100.0%
What blobs actually are — and what they're not
Blobs are a NEW data type added to Ethereum blocks (alongside transactions). Each blob is 128 KB. Each block can contain 0-6 blobs (target: 3). Crucially, blobs are STORED FOR ~18 DAYS THEN PRUNED — they don't bloat the chain forever like calldata does. This is the breakthrough: rollups need data temporarily (long enough to dispute/verify), not forever. Blobs satisfy that requirement at 5-100× lower cost than calldata storage.
Blob fee market — independent from regular gas
Just like EIP-1559's base fee, blob_base_fee adjusts each block based on blob utilization. Target: 3 blobs/block. Adjustment: blob_base_fee × exp((blobs_used - 3) / 3.34). When 6 blobs are used (max), fee multiplies by ~e^0.9 = 2.46× per block. When 0 blobs are used, fee divides by ~e^0.9 = 2.46× per block. This separate market means blob fees and regular gas fees move independently — high regular activity doesn't drag blob fees up.
Rollup batch economics — why L2 fees crashed
Pre-4844 (e.g., Arbitrum in January 2024): A rollup batch costs ~1M gas of calldata at 30 gwei = ~$100. That cost gets split among the ~5000 L2 transactions in the batch = $0.02 per user. But during high L1 gas (200 gwei), it's $700 / 5000 = $0.14 per user. Post-4844: A blob at 1 wei costs $0.0005 total. Even at 1000 gwei blob_base_fee (extreme), still $0.50 per blob = $0.0001 per user. Result: L2 fees are now within 1-2× of L1 transaction cost over the whole stack, and the variability dropped to near zero.
Why blob target is only 3 — and what's next
Higher blob target = more rollup throughput, but more bandwidth/storage burden on all Ethereum nodes. Target = 3 (max 6) was a conservative starting point. The roadmap: increase target to 6 (max 12), then to 32+ as PeerDAS launches (Peer-to-peer Data Availability Sampling — nodes only download a sample of each blob, verify correctness via cryptography). Full Danksharding (years away) targets 64 blobs/block = 8 MB/block of data, enough for 100k+ TPS across all rollups.
Frequently asked questions
+Why is blob_base_fee usually 1 wei (basically free)?
Because blob demand has been below target most of the time since launch. Even popular L2s (Arbitrum, Base, Optimism) collectively use about 3 blobs/block average — right at target. When usage drops below target, blob_base_fee decreases multiplicatively each block, quickly hitting the floor of 1 wei. Only during rare high-demand bursts (NFT mints on L2s, mass DEX activity) does blob_base_fee rise meaningfully.
+Do all L2s use blobs now?
Most major rollups switched within days/weeks of Dencun: Arbitrum (March 14), Optimism (March 14), Base (March 14), zkSync (April), Starknet (June), Scroll (April). They all saw immediate L2 fee reductions. Exceptions: some sidechains and chains using validium / volition data availability don't use blobs (they post elsewhere). Native L1 transactions still use regular calldata; only rollup batch data moved to blobs.
+If I'm an L2 user, do I need to do anything to benefit?
No — L2 user fees automatically dropped when L2s switched to blobs. You'll see the lower fees in your wallet immediately. If you're an L2 DAPP DEVELOPER, your gas costs for your dapp's transactions also dropped automatically (the L2 rollup eats the blob cost; you pay the L2 in L2 ETH at L2 gas prices).
+What happens when ALL 6 blob slots fill consistently?
Blob_base_fee rises exponentially each block (2.46× per max-block) until demand throttles back. In extreme scenarios this could push blob fees high enough that some L2s temporarily pause batches or use alternative data availability. Has happened a few times since launch (during peak NFT mints, popular L2 mass events) but typically resolves within minutes as the fee market clears.
+When does the blob target increase?
Pectra upgrade (May 2025) increased target to 6 blobs/block (max 9). Future upgrades (Fusaka, Glamsterdam) plan to increase further as PeerDAS rolls out. The roadmap targets 32+ blobs/block by 2027-2028, allowing significantly more rollup throughput.